Information elicitation for aggregate demand prediction with costly forecasting

This paper presents a multiple newsvendor-type purchasing problem where demand forecasts of a number of individual consumer agents can be generated at a price. Firstly, we derive the optimal solution for the model. Next, an information elicitation mechanism is presented that results in the optimal solution despite the autonomous, self-interested participants and the information asymmetry in between consumers and the supplier. Specifically, the incentive compatibility, efficiency, individual rationality and budget balance properties of the mechanism are proved and also illustrated by several numerical experiments.